Various cloud conditions can present risks to aircraft when traveling through them. If the temperature of a cloud atmosphere is below the freezing point for water, water droplets can become super-cooled liquid droplets. These super-cooled liquid droplets can then undergo a liquid-to-solid phase change upon impact with an aircraft surface. Ice accretes at different surface regions for different sizes of the super-cooled liquid droplets in the cloud atmosphere. Thus, characterizing the sizes of super-cooled water droplets in a cloud atmosphere can facilitate prediction of surface regions where ice will accrete as well as providing alerts of potentially dangerous conditions to a pilot.
Super-cooled small water droplets tend to form ice only on leading edges of an aircraft's exterior surface. Super-cooled Large water Droplets (SLDs), however, can strike the leading edge of a wing and run back past the icing protection systems, or can traverse airflow vectors and strike surfaces aft of these leading edges. Ice that forms on unprotected surface regions can severely alter the aerodynamics of the aircraft. Such ice accretion may cause aircraft stall or result in unpredictable aircraft control variation that might lead to flight issues. When in a cloud, ice can form on control surfaces and/or lift surfaces.
Not every cloud, however, has a significant SLD population. Different clouds and different atmospheric conditions may be accompanied by various water droplet size distributions, different ice/liquid ratios, etc., some of which may be entirely safe to an aircraft, while others may not be safe. Such water droplet size distributions and ice/liquid ratios may be measured as cloud metrics using various types of instruments.
Some aircraft are equipped with Light Detection and Ranging (LIDAR) systems to measure cloud metrics. Such systems can characterize clouds that have water droplets that have a size distribution having a single mode. Either the mean droplet size or the mode droplet size can be calculated by inversion of a backscatter signal using such systems. These systems can also calculate the density of water droplets for such mono-modal distributions.
Multi-modal distributions of water droplet sizes, however, can be difficult to characterize. Such multi-modal distributions may occur, for example, when cumulus clouds drop drizzle or rain into a lower stratiform cloud deck, creating icing conditions. For droplet size distributions having a dominant mode and a secondary mode (e.g. large distribution of relatively small water droplets plus a small distribution of large water droplets), it can be difficult to detect the anomalous amounts of large water droplets in the secondary mode.
LIDAR systems project pulses of a collimated laser beam into the cloud atmosphere and then sense the signal backscattered by the cloud atmosphere. The collimated laser beam samples a relatively small volume of the cloud, due to the collimated beam having a small field of view (e.g., 4 mrad of divergence is not atypical). Sampling such a small cloud volume can result in the beam encountering few, if any of the SLDs of a secondary distribution.
Depending on the size and density of the SLDs in the secondary distribution, the backscatter signal can appear as scintillation spikes superimposed on an otherwise smooth continuous range-resolved backscatter signal characteristic of the primary distribution. The size and frequency of occurrence of the scintillation spikes depends on the sizes of the SLDs and on the volume of space probed by the collimated laser beam.
Unlike the smooth range-resolved backscatter signal from the primary distribution, backscatter signals from small distributions of large droplet can have randomly occurring scintillation spikes. Averaging of such backscatter signals over multiple laser pulses, while boosting the signal-to-noise ratio of the sparse droplet contribution, can cause the sporadic scintillation spikes for the sparse large droplet distribution to be attenuated, and perhaps even fall below a noise floor. Thus, the SLDs, which can be hazardous to aircraft, may not be sensed.
Mono-modal distributions of SLDs can also be problematic, if the density of SLDs is small. Again, the backscatter signal can be characterized by scintillation spikes randomly located in the temporal stream of backscatter signals. Averaging of such backscatter signals can result in a signal amplitude that is small. Such a small signal may even fall below an instrument noise floor. Measurement techniques and instruments, which can more accurately characterize water droplet distributions, are needed.